Skip to main content
Ctrl+K

MASE 0.0.1 documentation

Overview

  • Installation
    • Getting Started using Conda
    • Getting Started using Docker
    • Getting Started using Nix
    • Additional Instructions for Imperial College Students
  • Quickstart
  • Tutorials
    • Tutorial 1: Introduction to the Mase IR, MaseGraph and Torch FX passes
    • Tutorial 2: Finetuning Bert for Sequence Classification using a LoRA adapter
    • Tutorial 3: Running Quantization-Aware Training (QAT) on Bert
    • Tutorial 4: Unstructured Pruning on Bert
    • Tutorial 5: Neural Architecture Search (NAS) with Mase and Optuna
    • Tutorial 6: Mixed Precision Quantization Search with Mase and Optuna
    • Advanced: TensorRT Quantization Tutorial
    • Advanced: ONNX Runtime Tutorial
    • Advanced: Using Mase CLI
    • Developer: Guide on how to add a new model into Chop
    • Developer: How to write documentations in MASE
    • Developer: How to extend search
  • Coding Style Specifications
    • Python Coding Style Specifications

Chop API

  • Chop Documentation
    • chop.actions
    • chop.datasets
    • chop.distributed
    • chop.ir
    • chop.models
    • chop.nn
    • chop.nn.quantized
      • chop.nn.quantized.functional
      • chop.nn.quantized.modules
    • chop.passes
      • chop.passes.module
        • chop.passes.module.transform.quantize
        • chop.passes.module.transform.quantize
      • chop.passes.graph
        • chop.passes.graph.analysis.add_metadata
        • chop.passes.graph.analysis.autosharding
        • chop.passes.graph.analysis.init_metadata
        • chop.passes.graph.analysis.report
        • chop.passes.graph.analysis.statistical_profiler.profile_statistics
        • chop.passes.graph.analysis.verify.verify
        • chop.passes.graph.calculate_avg_bits_mg_analysis_pass
        • chop.passes.graph.pruning
        • chop.passes.graph.analysis.runtime
        • chop.passes.transform.pruning
        • chop.passes.transform.quantize
        • chop.passes.transform.utils
        • chop.passes.transform.tensorrt
        • chop.passes.interface.save_and_load
        • chop.passes.interface.tensorrt
        • chop.passes.interface.onnxrt
    • chop.pipelines
    • chop.tools

Advanced Deep Learning Systems

  • Advanced Deep Learning Systems: 2024/2025
    • Lab 0: Introduction to Mase
    • Lab 1: Model Compression (Quantization and Pruning)
    • Lab 2: Neural Architecture Search
    • Lab 3: Mixed Precision Search
    • Lab 4 (Software Stream) Performance Engineering
    • ADLS Docker Environment Setup
  • Advanced Deep Learning Systems: 2023/2024
    • Lab 1 for Advanced Deep Learning Systems (ADLS)
    • Lab 2 for Advanced Deep Learning Systems (ADLS)
    • Lab 3 for Advanced Deep Learning Systems (ADLS)
    • Lab 4 (Software Stream) for Advanced Deep Learning Systems (ADLS)
    • ADLS Docker Environment Setup
  • .rst

Chop Documentation

Chop Documentation#

Chop is the software stack of Mase.

Chop

Chop provides the CLI, training and testing actions, graph passes, datasets, and model tooling used throughout Mase.

Chop API

  • chop.actions
    • chop.actions.train
    • chop.actions.test
    • chop.actions.transform
    • chop.actions.search
  • chop.datasets
    • chop.dataset.nerf
    • chop.dataset.nlp
    • chop.dataset.physical
    • chop.dataset.vision
  • chop.distributed
  • chop.ir
    • chop.ir.graph
  • chop.models
  • chop.nn
    • chop.nn.functional
    • chop.nn.modules
  • chop.nn.quantized
    • chop.nn.quantized.functional
    • chop.nn.quantized.modules
  • chop.passes
    • chop.passes.module
    • chop.passes.graph
  • chop.pipelines
    • chop.pipelines.auto_pipeline
  • chop.tools
    • chop.tools.check_dependency
    • chop.tools.checkpoint_load
    • chop.tools.config_load
    • chop.tools.get_input
    • chop.tools.logger
    • chop.tools.onnx_operators
    • chop.tools.registry
    • chop.tools.utils

previous

Python Coding Style Specifications

next

chop.actions

By DeepWok

© Copyright 2023, DeepWok.